Human-Centered Artificial Intelligence

Human-centered artificial intelligence is a new paradigm that is now the focus of the research and development activities of both large companies and prestigious international research centers in the U.S. and Europe. The underlying idea is to shift the focus of Artificial Intelligence from "stand-alone" applications aimed at replacing humans in intelligent tasks to interactive applications in which human and machine intelligence work together to overcome the limitations of both.

The overall goal of this master's degree program is to offer a significantly interdisciplinary context to train new professionals capable of accompanying the widespread diffusion of Artificial Intelligence, enabling the reasonable and responsible integration of new technologies into the human context in which they are to be used. This integration aims to solve complex problems involving a constellation of non-technical variables: strategic goals, moral values, legal constraints, cognitive biases, and other psychological and social factors. From this perspective, the input of human agents becomes an integral part of an Artificial Intelligence system, and Artificial Intelligence itself becomes a set of sophisticated technologies to enhance the intelligence of human agents by expanding their cognitive capabilities. Hence, there is an essential need for strongly interdisciplinary skills, both theoretical and empirical ones, to meaningfully and responsibly guide this process of integrating new technologies into the real-world contexts.

The aim is to train bridging figures between the hard-skills of experienced developers and the soft-skills needed to integrate AI applications into the human context in which they are deployed.

Graduates in Human-Centered AI will possess:
- significant education in the disciplines characterizing the interaction between human cognition and AI;
- a thorough understanding of the most advanced methods of data collection, analysis and decision-making (machine learning);
- a thorough understanding of the theoretical, technical and cognitive aspects of human-computer interfaces;
- the ability to design models and interventions for the reorganization of interfaces between humans and AI systems;
- the ability to independently conduct research activities in the field of artificial intelligence;
- the ability to use fluently, in written and oral form, at least one language of the European Union besides Italian, with reference also to the terminology of the field;
- a thorough knowledge of a theoretical and operational nature on communication and decision-making processes based on the use of artificial intelligence systems;
- knowledge of the principles and the main methodologies of AI at a level adequate to interact fruitfully with computer scientists and connect them with domain experts in a multi-disciplinary context;
- a thorough understanding of the non-technical aspects- e.g. ethical issues, legal constraints, cognitive aspects, philosophical foundations, neuroscientific foundations associated with the use of AI technologies to support, not replace, humans and their activities;
- familiarity with the main applications of AI in the work context (business, health care, legal) and with the tools that enable informed and transparent interactions between humans and machines;
- decision-making abilities in highly complex situations.

Specific objectives will be formulated according to the curricula in which the course is organised. Since this is an interdisciplinary degree, which admits graduates from different backgrounds, these objectives will be achieved:
a) by including in the curriculum alternative courses that allow students to integrate their previously acquired knowledge according to the degree course of origin and the exams taken,
b) by proposing personalised study plans to guide students in their choices,
c) by proposing advanced foundations courses for the characterising subjects that are essential to the achievement of the training objectives, the purpose of which is to provide, in the initial part, a summary of the basic knowledge needed to acquire more advanced content,
d) by setting up a structured tutoring service to facilitate students with different backgrounds using these courses.

Starting from a broad common core, the course will be divided into three curricula. The common core will consist of characterising subjects belonging to the following areas:
1) philosophical and linguistic disciplines (with the addition of the areas IUS-20 and IUS-08), to acquire knowledge and competences of logical, epistemological and ethical-legal type
2) psychological disciplines, to acquire knowledge and competences on human-computer interaction and on the role of AI in decision-making processes
3) psychobiological and neuroscience disciplines, to acquire knowledge and skills relating to cognitive functions and their neural bases
4) mathematical, computer and engineering disciplines, to acquire knowledge and skills relating to development and analysis of machine learning models, algorithms and programming, for knowledge representation and reasoning, and natural language processing.

Laboratories aimed at acquiring further knowledge and skills in computing are also part of the common core. The three curricula aim to provide a more specific preparation in relation to three main contexts:

A) the general context of integrating AI applications in an organisation and planning for fruitful collaboration between humans and machines, taking into account the psychological and social component of this interaction. This curriculum will provide:
- additional knowledge and skills in the field of AI, obtained through teaching in the fields of mathematics, computer science and engineering, as well as additional computer laboratories;
- knowledge and skills relating to the psycho-social and legal aspects of working in complex teams (made up of human beings with different skills and machines) and the impact of AI on the organisation of work, obtained through teaching in the field of psychology and related and complementary disciplines, with reference to sociology and anthropology.

B) The context of clinical and theoretical neuroscience. This curriculum will provide
- further knowledge and skills in the field of neural bases of brain processes for the development of AI-based, multiscale and bio-inspired neural models and for the handling of human-machine interface neural signals;
- knowledge and skills relating to the application of AI algorithms in the field of clinical neuroscience, in order to promote the diagnostic and therapeutic/rehabilitation process in the direction of precision and personalised medicine.

C) The legal (domestic and European) as well as ethical context of AI applications in a public or private organisation. This curriculum will provide:
- further knowledge and skills in the field of AI, adopting a multidisciplinary approach that enables the combination within the same teaching course (and laboratories) of the mathematical, computer and engineering disciplines relating to a specific field of application (judicial, public administration, tax, labour relations etc.), with the respective and specific legal issues;
- knowledge and skills relating to the general ethical-legal aspects associated with AI applications, such as profiles relating to fundamental and human rights, data protection and data collection, civil and criminal liability, protection of intellectual property, communication, transparency.

All curricula guarantee, within the framework of the characterising disciplines, a minimum of 12 ECTS of computer science teaching in the first year, aimed at consolidating, or providing, if necessary, fundamental knowledge and skills in this field. The laboratory activities also provide, for all students, the acquisition of at least a further 9 ECTS in activities useful for acquiring computer skills.